From owner-freebsd-isp  Tue Nov 19 14:38:54 1996
Return-Path: owner-isp
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id OAA11768
          for isp-outgoing; Tue, 19 Nov 1996 14:38:54 -0800 (PST)
Received: from red.jnx.com (red.jnx.com [208.197.169.254])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id OAA11761
          for <isp@freebsd.org>; Tue, 19 Nov 1996 14:38:45 -0800 (PST)
Received: from chimp.jnx.com (chimp.jnx.com [208.197.169.246]) by red.jnx.com (8.8.3/8.8.3) with ESMTP id OAA12890; Tue, 19 Nov 1996 14:38:12 -0800 (PST)
Received: (from tli@localhost) by chimp.jnx.com (8.7.6/8.7.3) id OAA16243; Tue, 19 Nov 1996 14:38:02 -0800 (PST)
Date: Tue, 19 Nov 1996 14:38:02 -0800 (PST)
Message-Id: <199611192238.OAA16243@chimp.jnx.com>
From: Tony Li <tli@jnx.com>
To: dennis@etinc.com
CC: isp@freebsd.org
In-reply-to: <199611191516.KAA06950@etinc.com> (dennis@etinc.com)
Subject: Re: changed to: Frac T3?
Sender: owner-isp@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk


   I think that tony has been working on scarce resouce machines for too
   long.  

Well, that's certainly true.  However, PC's don't have significantly more
resources, so...  

   You buy time with buffer space, and buffer space in a freebsd
   enviroment is rather abundant.  You've got hardware buffering, and OS
   buffering, all which need to compensate for the time needed to service
   the interrupts.

Sorry, this just false.  Buffering allows you to trade rate mismatches for
latency.  Unfortunately, you have a hard latency limitation.  So adding
4GBytes of buffering to your FreeBSD box does NOT make the routing protocol
more stable.

   At T1 there is no issue, because you can buffer several full timeouts
   worth of data.

Yes, but so what?  If you queue up my protocol packets then either I can't
get at them for several seconds.  Alternately, if this is output buffering
and my protocol packets don't get out of the box for several seconds, I'm
equally toast.

   I think building a freebsd box with 1 100Mbs ethernet and 1very high
   speed interface is fairly easy...because you can control the process
   flow.

I agree with the goal and the conclusion.  I still don't believe that
you've got enough process level control that you can also make the box a
Web server, say and not endanger the protocols.

Tony